Image Classification on Oxford-IIIT Pets (test) (Reporting Baseline Delta)
92.29Top-1 AccuracyComMem
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ComMemBackbone=ViT-B/162026.06 | 92.29 | — | |
| DMN-ZSBackbone=ViT-B/162026.06 | 92.04 | — | |
| DPEBackbone=ViT-B/162026.06 | 91.14 | — | |
| CoOpBackbone=ViT-B/162026.06 | 89.14 | — | |
| TDABackbone=ViT-B/162026.06 | 88.63 | — | |
| DynaPromptBackbone=ViT-B/162026.06 | 88.28 | — | |
| CLIPBackbone=ViT-B/162026.06 | 88.25 | — | |
| DiffTPTBackbone=ViT-B/162026.06 | 88.22 | — | |
| TPTBackbone=ViT-B/162026.06 | 87.79 | — | |
| TPSBackbone=ViT-B/162026.06 | 87.35 | — | |
| CoOpBackbone=ResNet-502026.06 | 87 | — | |
| EnsembleBackbone=ViT-B/162026.06 | 86.92 | — | |
| DMN-ZSBackbone=ResNet-502026.06 | 86.78 | — | |
| ComMemBackbone=ResNet-502026.06 | 86.59 | — | |
| TDABackbone=ResNet-502026.06 | 86.18 | — | |
| DPEBackbone=ResNet-502026.06 | 85.97 | — | |
| TPTBackbone=ResNet-502026.06 | 84.49 | — | |
| CLIPBackbone=ResNet-502026.06 | 83.57 | — | |
| DiffTPTBackbone=ResNet-502026.06 | 83.4 | — | |
| DynaPromptBackbone=ResNet-502026.06 | 83.07 | — | |
| EnsembleBackbone=ResNet-502026.06 | 82.97 | — | |
| Baseline (No Accel.)Acceleration Type=None2026.03 | 81.03 | — | |
| BiGainTDAcceleration Type=Token Downsampling, FLOPs Reduction=14.2%2026.03 | 79.9 | 1.13 | |
| ToDoAcceleration Type=Token Downsampling, FLOPs Reduction=14.2%2026.03 | 79.15 | 1.88 | |
| BiGainTMAcceleration Type=Token Merging/Pruning, FLOPs Reduction=10%2026.03 | 78.38 | 2.65 | |
| MosaicDiffAcceleration Type=Model Pruning, FLOPs Reduction=10%2026.03 | 77.38 | 3.65 | |
| DiP-GOAcceleration Type=Model Pruning, FLOPs Reduction=10%2026.03 | 76.53 | 4.5 | |
| ToMeAcceleration Type=Token Merging/Pruning, FLOPs Reduction=10%2026.03 | 72.96 | 8.07 | |
| SiToAcceleration Type=Token Merging/Pruning, FLOPs Reduction=7%2026.03 | 68.84 | 12.19 |